SqueezeViX-Net with SOAE: A Prevailing Deep Learning Framework for Accurate Pneumonia Classification using X-Ray and...

Area of Science:

• Medical Imaging Analysis
• Artificial Intelligence in Healthcare
• Deep Learning for Disease Detection

Background:

• Pneumonia is a severe respiratory illness requiring timely and accurate diagnosis for effective treatment.
• Delayed or incorrect diagnosis of pneumonia increases mortality rates, especially in vulnerable populations.
• Accurate pneumonia classification is crucial for appropriate clinical management and patient outcomes.

Purpose of the Study:

• To introduce SqueezeViX-Net, a deep learning framework for precise pneumonia classification.
• To enhance model stability and suitability using a Self-Optimized Adaptive Enhancement (SOAE) method.
• To evaluate the performance of SqueezeViX-Net on diverse medical imaging datasets.

Main Methods:

• Developed SqueezeViX-Net, a deep learning model tailored for pneumonia classification.
• Implemented a Self-Optimized Adaptive Enhancement (SOAE) technique to dynamically adjust dropout rates during training.
• Validated the model using extensive X-ray and CT image datasets from Kaggle repositories.

Main Results:

• SqueezeViX-Net demonstrated superior performance compared to established architectures like DenseNet-121, ResNet-152V2, and EfficientNet-B7.
• The model achieved higher accuracy, precision, recall, and F1-score metrics.
• Validation across varied pneumonia datasets, including CT and X-ray images, confirmed its robustness and ability to handle different imaging modalities.

Conclusions:

• SqueezeViX-Net, incorporating SOAE technology, offers an advanced framework for specific pneumonia identification in clinical settings.
• The model's dynamic learning capabilities and high precision present significant potential for medical professionals.
• This AI tool can contribute to improved patient treatment strategies and outcomes in pneumonia care.